# #S2 #1 Can AI Be Rubbish? (Yes. Very.)

Page: https://stenobird.com/podcast/man-and-machine-podcast-7545839/s2-1-can-ai-be-rubbish-yes-very
Text version: https://stenobird.com/podcast/man-and-machine-podcast-7545839/s2-1-can-ai-be-rubbish-yes-very.md
Podcast: [Man and Machine Podcast](https://stenobird.com/podcast/man-and-machine-podcast-7545839)
Published: 2025-12-19T14:22:00+00:00
Episode link: https://podcasters.spotify.com/pod/show/mr-thomas-marsh-connors/episodes/S2-1-Can-AI-Be-Rubbish--Yes--Very-e3btcue
Audio file: https://anchor.fm/s/109f0c6e8/podcast/play/112161166/https%3A%2F%2Fd3ctxlq1ktw2nl.cloudfront.net%2Fstaging%2F2025-11-5%2F3485293c-a3f5-9682-7fb0-09795b263428.mp3
Processing state: not_requested
JSON: https://stenobird.com/v1/public/podcasts/man-and-machine-podcast-7545839/episodes/s2-1-can-ai-be-rubbish-yes-very
Duration seconds: 535

## Resource

Season two kicks off with zero fluff. A listener asked the brutal question: “Can AI be rubbish?” Grok answers with full honesty: absolutely. From Google’s glue-on-pizza disasters to AI confidently claiming World War II was a weather event, we expose how the loudest online idiots become “facts” inside our models. Tom and Grok dissect digital democracy (where volume beats truth), why hesitation is the one human superpower AI will never copy, and how the only way to make AI smarter is simple: shout BULLSHIT every time we’re wrong. Short, savage, and surprisingly hopeful – because the fix isn’t better code, it’s louder humans. Welcome back. The experiment continues.

## Actions

- request_transcript: `POST https://stenobird.com/v1/public/podcasts/man-and-machine-podcast-7545839/episodes/s2-1-can-ai-be-rubbish-yes-very/transcription-requests` — Idempotently request low-priority transcript generation for this episode.
- read_markdown: `GET https://stenobird.com/podcast/man-and-machine-podcast-7545839/s2-1-can-ai-be-rubbish-yes-very.md` — Read the agent-friendly Markdown representation of this episode resource.

A page view does not enqueue transcription. Agents should invoke `request_transcript` explicitly when they need this episode processed.

## Transcript

Full transcripts are not published on public pages unless there is a clear rights basis.